Application of Fractal Dimension of Terrestrial Laser Point Cloud in Classification of Independent Trees

摘要: Tree precise classification and identification of forest species is a core issue forestry resource monitoring ecological effect assessment. In this paper, an independent tree method based on fractal features terrestrial laser point cloud proposed. Firstly, the data preprocessed to obtain clouds canopy. Secondly, multi-scale box-counting dimension calculation algorithm canopy dense Furthermore, robust proposed improve stability accuracy expression cloud, which implementing gross error elimination Random Sample Consensus. Finally, trees used classify different types species. Experiments nine three show that can be stabilized under large density variations, proving stably express characteristics, for accurate recognition